A Graph-oriented Measure of Rule Adaptivity and Flexibility

L.Y. Tan and J.C. Tay (Singapore)

Keywords

Visualization, Rulebased Program, Software Complexity.

Abstract

A common issue with rule-based systems is the ease with which rules in the system may be changed. As rules are added, modified and deleted, the relationships between the rules become increasingly complex. Beyond which, the logic behind the asynchronous nature of these rule executions will prevent users from being able to predict the effects of intended changes, possibly leading to an increasing unstable and unmaintainable system. This paper describes an intuitive method of visualization that allows developers to understand the control dependency between rules and the potential rule activation paths that may occur during runtime. Based on this visualization method, a software metric is proposed that measures the strength of individual rules in effecting rule activations, which may be used in flexibility and reusability studies of the rule-based system. We evaluate our results through visualization and complexity measures of rule-based versions of Warshall and Dijkstra-Lee shortest distance algorithms.

Important Links:



Go Back